Perceived Peer Behavior and the Timing of Sexual Debut in Rwanda: A Survival Analysis of Youth Data

This paper uses survival analysis methods to assess the relationship between perceived peer behaviour and the timing of sexual debut among Rwandan youth. The perception that peers are sexually active is associated with increased likelihood of early sexual experience. The relationship between perceived peer behaviour and the timing of first sex is stronger in young adulthood than in late adolescence for both sexes, and in female youth, the presence of the father appears to weaken the impact of peer behaviour.

Background

UNAIDS has estimated that about 11.8 million youth aged 15–24 years are currently living with HIV in part because of early, premarital sexual relations, multiple sexual partnerships, and inconsistent condoms use. The goal of this study is to help identify the potentially modifiable factors that affect sexual behaviours and susceptibility to HIV infection among youth, so that behaviour change programmes can be targeted for greater effect.

The author cites studies that suggest that peers play a strong role in risk-taking among youth. What their peers are doing or what they perceive them to be doing influences youth behaviour in a variety of settings. The analyses presented in this paper were done against the background of an ideation model, "an analytic framework that incorporates constructs from leading behaviour change theories." Ideation is defined by the author as "ways of thinking that can be diffused through communication and social interactions." The main assumption behind the ideation model is that the most proximate determinants of specific behaviours are the factors reflecting a person’s ways of thinking.

Methodology

The model predictors are divided into three categories to make up the three dimensions of ideation - the basis for measurement and assessment of the link to a variety of reproductive health behaviours. These three dimensions consist of cognitive, emotional, and social interaction factors. Peer behaviour and perceived peer behaviour are identified as two of the variables indicating the cognitive dimensions of ideation. The author's working hypothesis is that a strong and positive relationship exists between perceived peer sexual behaviour and early sexual experimentation. The respondents who perceive widespread sexual activity among their friends will be more likely to experience an early sexual debut than their peers who do not.

This study is based on survey data collected in May 2001 among Rwandan youth (n=1327) in four provinces - Kigali Ville, Butare, Kibuye, and Umutara. The sample for the survey is representative of male and female youth aged 15-24 years in the study area. Interviews were conducted using a structured questionnaire that inquired about sex-related attitudes and behaviour, self-concept, HIV/AIDS knowledge and attitudes, substance use, and media exposure. This paper uses survival analysis methods - life table analysis and Cox proportional hazards method - to identify the patterns and correlates of the age at first sex (a full discussion of the background and characteristics of these methods are discussed in the paper).

In addition to several socio-demographic independent variables, the author tests the following ideation related independent variables:

• Perceived peer sexual behaviour
• Self-efficacy to refuse sex
• Self-esteem
• Perceived social support for premarital sexual abstinence

Results

Analyses show that the median age at first sex is 20.9 years for male youth and 20.1 years for female youth and the author notes that these findings are comparable to what was found by the Rwanda 2000 Demographic and Health Survey (DHS). The proportion that is yet to experience sex is much lower among those who perceive the majority of their peers to be sexually active than among those who do not hold such a perception. There are significant differences in the timing of first sex by perceived peer behaviour among both male and female youth, though the author admits that the observed tendency may be due to possible greater underreporting of sexual activities by females. Among female youth, the perception that all or most friends are sexually active does not appear to result in increased sexual activity prior to age 15, whereas among male youth the perception is associated greater sexual activity from age 10 and up.

The results of the Cox Proportional Hazard Model provide the basis for ascertaining the relationship between perceived peer behaviour and the timing of first sex. The predominant, statistically significant (sig. varies from p 0.05 to p 0.0001) hazard ratio factors relating to timing of first sex are as follows (using the authors second model that includes perceived peer behaviour): Baseline = 1:

Males

* Religion = Moslem (2.49)

* Residence = Urban (1.71)

* School = Enrolled (0.55)

Ideation Variables

* Belief that sex should only come after marriage (0.57)

* Perceived peer behaviour (1.94)

Females

* Living arrangement = only lived with father as child (1.92)

* Living arrangement = never lived with father (2.38)

* Alcohol = ever used alcohol (1.91)

Ideation Variables

* Perceived self-efficacy to refuse sex with someone loved truly (0.49)

* Belief that sex should only come after marriage (0.66)

* Perceived peer behaviour (2.57)

Conclusions

The author notes that there is some evidence of underreporting of sexual activity, particularly in the younger age cohorts, thus the age of sexual debut is probably underestimated. Regardless of this problem, the author believes that the evidence supports the hypothesis that perceived peer behaviour influences early sexual debut, more so for female youth than for male youth. Two possible explanations are posited: youth select friends that conform to their own behaviours, or youth adopt the behaviour of their admired peers. Further, panel based studies would be needed to determine this relationship. Other findings include clear evidence of the important role of fathers in dissuading female youth from early sexual experiences, though not for males, suggesting that gender bias plays a role. The belief that sex should only come after marriage was also a major determinant, as was being Moslem for boys (religion was also a strong factor for girls before the peer perception factor was included). The author concludes by making a series of recommendations with regards to behaviour interventions, including the necessity of peer-to-peer communication activities, life skill training such as sexual negotiation, and programming that is gender and age specific.